Effects Of Complex Radiative And Convective Environments On The Thermal Biology Of The White-Crowned Sparrow (Zonotrichia Leucophrys Gambelii)

The maintenance of high body temperatures is a costly enterprise for small endotherms and in many species may account for a major portion of their daily energy expenditure. The basal metabolic and thermoregulatory costs of small birds, for example, typically account for 40–60 % of their total daily energy expenditure. These rates of energy expenditure are strongly influenced by characteristics of the physical environment or the ‘microclimates’ the animal occupies. Surprisingly, of the physical environmental factors that define an animal’s microclimate (e.g. air temperature, solar radiation and wind speed), only the effects of air temperature are well studied (for a review, see Calder and King, 1974). The effects of wind speed on avian thermoregulatory metabolism have also received some attention, and convective heat transfer clearly plays a significant role in determining the heat balance and thermoregulatory costs in small birds (for a review, see Goldstein, 1984; see also Bakken et al., 1991; Buttemer, 1981; Webb and Rogers, 1988; Webster and Weathers, 1988; Wolf and Walsberg, 1996).

Solar radiation represents a potentially overwhelming heat load to small animals and may therefore significantly influence thermoregulatory metabolism; under clear skies, it commonly exceeds 900–1000 W m⁻² on a plane perpendicular to the solar beam. To date, only a handful of studies have focused on the effects of solar radiation on heat balance and thermoregulation in small birds (DeJong, 1976; Hamilton and Heppner, 1967; Heppner, 1970; Lustick, 1969, 1971; Lustick et al., 1970), all these studies being carried out in the absence of wind (e.g. under free-convective conditions).

Taken together, these analyses provide important insight into the thermoregulatory responses of small birds to challenges from their thermal environment, but impart an incomplete picture of how animals respond physiologically to complex natural environments. Especially lacking is an understanding of the interactive effects between solar radiation and forced convection on the heat balance and thermoregulatory costs of wild birds; empirical data are available from only a single study of the desert-dwelling verdin, Auriparus flaviceps (Wolf and Walsberg, 1996). These data suggest that the convective environment has a significant influence on radiative heat loads transferred from the animal’s plumage surface to its skin. Unfortunately, it is difficult generalize how forced convection affects radiative heat transfer from a single study. In addition, the available data set may be not indicative of a generalized response because of the extremely small body size (approximately 6.5 g) of the verdin and its light insulation (e.g. plumage mass was 29 % lower than allometric predictions; Webster and Weathers, 1988).

Radiative heat flow from the plumage surface to the bird’s skin is affected by variations in wind speed because of alterations in boundary layer and plumage heat transfer resistance. In birds, incident radiation does not simply interact with feather elements at the plumage surface to produce heat, but may penetrate more deeply into the plumage. Thus, the heat load that the skin receives is influenced not only by such conspicuous factors as wind speed and coat color but also by more subtle properties such as the microstructure and optics of the feather elements. In mammals, such factors have been found to have an overwhelming effect on the heat load transferred to the skin (Walsberg and Wolf, 1995b). Finally, other factors, such as changes in peripheral circulation and pilo-or ptilo-erection in response to irradiation, may greatly modify the total thermal resistance of the animal’s coat. Unfortunately, it is a difficult task to integrate these factors and to estimate accurately, or to model in a simple manner, the effects of varying convective environments on the radiative heat loads experienced by animals that possess fibrous pelts or plumages.

In this investigation, we examine how complex radiative and convective environments affect thermoregulatory costs and heat transfer in a small bird. We focus on the following questions. (i) How does solar radiation and wind speed affect metabolic heat production in a small finch? (ii) To what degree does forced convection affect physiologically significant solar heat gain? Physiologically significant solar heat gain (SHG) is quantified as the decrease in metabolic heat production caused by the animal’s exposure to simulated solar radiation. Similarly, changes in convective heat loss are quantified as the variation in metabolic heat production caused by exposure to different wind speeds. (iv) How much intercepted radiation is utilized by the animal, as shown by concurrent reductions in metabolic rate associated with irradiation? (v) How is body temperature affected by differing convective and radiative regimes? We chose the white-crowned sparrow (Zonotrichia leucophrys gambelii) as a study animal because of the extensive background research on its thermoregulatory processes and physiology (e.g. DeJong, 1976; Ketterson and King, 1977; King, 1964; Mahoney and King, 1977; Paladino and King, 1984; Robinson et al., 1976).

White-crowned sparrows Zonotrichia leucophrys gambelii were captured with mist nets in Apache Junction, AZ, USA, during January 1995. The birds were transported to Arizona State University, Tempe, Arizona, where they were housed individually in BioQuip 0.2 m³ folding insect cages in a temperature-controlled environmental room. The room temperature was maintained at 24±0.5 °C and the photoperiod was held at 10 h:14 h L:D. Sparrows were maintained on an ad libitum diet of chick starter and wild bird seed. Water and grit were also freely available.

Metabolic measurements were made both in the presence and absence of simulated solar radiation with the sparrows placed in a closed-circuit wind tunnel, described by Walsberg and Wolf (1995b). This wind tunnel had an effective volume of 45 l, calculated following Bartholomew et al. (1981). A variable-speed direct-current blower circulated air through the chamber test section at wind speeds (u) ranging from 0.25 to 2.0m s⁻¹, as measured using a Thermonetics HWA-101 thermoanemometer that had been calibrated as described by Walsberg (1988). Vertical and horizontal louvers installed upstream of the test section were adjusted to ensure uniform airflow within the chamber. Wind speed varied less than 5.0 % vertically and horizontally except within 1 cm of the chamber walls. Turbulence intensity was 3 % or less at all wind speeds, as determined using a thermoanemometer with the signal output measured by a true-root-mean-square voltmeter (Beckman model 850) and computed using the method of Hinze (1959).

Simulated solar radiation was produced by a Spectral Energy Corp. series II solar simulator, which filters light produced by a 1 kW xenon arc lamp to simulate direct solar radiation at an air mass of 1. Irradiance within the test section of the chamber was maintained at 936±11 W m⁻² (mean ± S.D.) (high irradiance) and was measured using a LiCor LI200sz pyranometer that had been calibrated against an Oriel Corp. model 7080 pyroelectric radiometer. Simulated solar radiation passed through a 4.8 mm flint glass window installed in the top of the chamber test section and onto the animal. The flint glass blocked the intense ultraviolet radiation produced by the lamp, which could potentially have burned the eyes and skin of the experimental animal. Wind tunnel measurements in the absence of simulated solar radiation (low irradiance) were carried out under fluorescent room lighting at an irradiance of less than 3 W m⁻².

Chamber air temperature was maintained at 10±0.5 °C by placing it within a temperature-controlled room and by circulating water from a constant-temperature reservoir through heat-exchange coils located in the walls of the test section and in the airstream downwind of the test section. Chamber air temperature was measured using a 26 gauge type T thermocouple calibrated at two points, at 0 °C in a distilled water ice bath and at 35 °C in a water bath against a mercury thermometer traceable to National Bureau of Standards and recorded by a Campbell model CR21x datalogger. The long-wave radiant environment in the test section of the chamber was held nearly constant by painting the walls of the chamber flat black. Assuming an emissivity of 0.98 and calculating the long-wave irradiance using the Stefan–Boltzmann relationship, we estimated that long-wave emission varied less than 6 %.

A Puregas model CDA1112 air dryer/CO₂ scrubber system provided dry, CO₂-free air that ventilated the metabolic chamber at a rate of 12 028±65 ml min⁻¹ (mean ± S.D.). Incurrent flow was measured using an Omega FL 3404T-HRV rotameter, calibrated to ±1 % of full scale with a 5 l soap-bubble flowmeter. These flow rates allowed the gas mixture within the metabolic chamber to reach 99 % equilibrium in approximately 17 min, as calculated following Lasiewski et al. (1966). A sample of chamber effluent air was dried, by passing it through a column of anhydrous calcium sulphate at 150 ml min⁻¹, and then passed into a LiCor model 6252 infrared CO₂ analyzer that determined CO₂ concentration to ±1.0 p.p.m. The CO₂ analyzer was zeroed and spanned daily using dry CO₂-free air and a calibration gas known to contain 0.1570 % CO₂. Instrument signals were recorded on a Campbell model CR21x datalogger and averaged at 1 min intervals.

Metabolic measurements were made during the photophase within the closed-circuit wind tunnel, both in the presence and absence of simulated solar radiation. Measurements were made on 10 adult white-crowned sparrows (mean body mass ± S.D., 22.9±1.5 g) between January and May 1995. Sparrows were blindfolded during measurements, to minimize locomotor activity, with a 1 cm×3 cm black Lycra strip that covered their eyes and through which their bill protruded; an elastic thread stretched around the back of the skull held the blindfold in place. Birds were placed on a dowel perch in the center of the test section of the chamber 2.5 cm above the chamber floor and rested quietly for approximately 60 min at each wind speed (u=0.25, 0.5, 1.0 or 2.0 m s⁻¹) while CO₂ production was measured. The order in which animals were exposed to wind speed/irradiance combinations was randomized, and sparrows were exposed to only a single wind speed and irradiance combination on a single day. We report data for the minimum 1 min average values for CO₂ production measurements taken during the last 7 min of each 1 h run. All measurements were corrected to standard conditions (0 °C, 101 kPa) using equation 6.5 of Mclean and Tobin (1987) for flow measurements using rotameters. A Magnavox CCD camera surveillance system mounted above the chamber test section monitored the activity of each bird during the experiments; runs in which birds did not sit quietly on the perch were excluded from analyses.

The body temperatures of sparrows were obtained, within 1 min of the end of each run, after the bird’s removal from the metabolic chamber. An 18 gauge Physitemp Teflon sheathed thermocouple with a time constant of 0.1 s was inserted approximately 15 mm into the sparrow’s cloaca, until a slight withdrawal did not result in a decrease in temperature. Thermocouple temperatures were read using an Omega model HH23 digital thermocouple thermometer.

Respiratory quotients (RQ) and total evaporative water loss (TEWL) were measured in a separate series of experiments because the high flow rates used in the wind tunnel experiments precluded accurate measurements of oxygen consumption. Methods used are outlined below, and details of calibration procedures and of the apparatus used can be found in Walsberg and Wolf (1995a). Measurements were made between 09:00 and 15:00 h on recently fed (not post-absorptive) birds (N=12) held in the dark at 10 °C. Birds were placed in 2 l metabolic chambers and rested quietly on 1.3 cm² mesh hardware cloth above 2 cm of mineral oil used for the collection of fecal material. Dry, CO₂-free air flowed through the metabolic chamber at a rate of 505 ml min⁻¹. A sample of chamber effluent air passed over a Thunder scientific humidity evaluation module (model PC-2101) and was dried by passing it through a column of anhydrous calcium sulphate at 100 ml min⁻¹ and into a LiCor model 6252 CO₂ analyzer that determined CO₂ concentration to ±1.0 p.p.m. After leaving the CO₂ analyzer, efflux air was dried and scrubbed of CO₂ by passing it through a small column of anhydrous calcium sulphate and Ascarite, and oxygen concentration was determined using an Applied Electrochemistry S3a oxygen analyzer calibrated using the procedures of Walsberg and Wolf (1995a).

We report data for the minimum 1 min average values for evaporative water loss, CO₂ production and oxygen consumption on measurements taken during the last 7 min of each 1 h run. Instrument signals were recorded on a Campbell model CR21x datalogger that averaged the values for each minute. The activity of each bird was monitored during the experiments using a Magnavox CCD camera surveillance system. We calculated oxygen consumption rates using equation 2 of Hill (1972) and CO₂ production rates using equation 3 of Walsberg and Wolf (1995a). Measurements were corrected to standard conditions (0 °C, 101 kPa) using equation 6.5 of Mclean and Tobin (1987) for flow measurements using rotameters.

We computed metabolic heat production from using 20.9 J ml⁻¹ as the thermal equivalent of carbon dioxide. We divided these values by plumage surface area estimated from body mass (g) using the Meeh equation, as modified by Walsberg and King (1978), to obtain surface-area-specific heat production (H? m).

Values are presented as means ±95 % confidence intervals except where otherwise noted. Mean values for body temperature measurements taken in the presence and absence of simulated solar radiation were compared at each wind speed using a paired two-sample Student’s t-test using Microsoft Excel. Analyses of covariance and least squares regression were carried out using JMP statistical software. We did not use a repeated-measures analysis of covariance (ANCOVA) because only two birds were used in all eight wind speed/irradiance treatments. Thirteen birds were used during the course of the study.

Daytime RQ averaged 1.01±0.01 (N=12). On the basis of this value, the thermal equivalent of carbon dioxide produced is estimated as 20.9 kJ l⁻¹ (Brody, 1945). Total evaporative water loss averaged 9.3±0.5 mg g⁻¹ h⁻¹ (N=12), and latent heat loss averaged 22.8 W m⁻², calculated using a value of 2.42 kJ g⁻¹ H₂O.

Fig. 1 presents T_b as a function of wind speed and irradiance. We found no correlation between T_b and wind speed, so we pooled T_b data for all wind speeds at a given irradiance level. T_b differed significantly with irradiance level at all wind speeds tested except 1.0 m s⁻¹ (Fig. 1). At T_a=10 °C, and at the low level of irradiance, T_b averaged 42.5±0.4 °C; at the high level of irradiance, T_b averaged 41.8±0.3 °C.

We examined differences in the relationships between metabolic heat production and wind speed at our two experimental irradiances (Fig. 2) using ANCOVA. Metabolic heat production increased linearly with the square root of wind speed (r²=0.67 for low irradiance and r²=0.75 for high irradiance; P<0.0001). Using the square root of wind speed provided a better linear fit than using untransformed wind speed, as noted by Webster and Weathers (1988). Between 0.25 and 2.0 m s⁻¹, metabolic heat production increased by 30 and 51 % at the low and high irradiances, respectively. Metabolic heat production H?_m as a function of the square root of wind speed u for low and high irradiance levels can be described by the following linear model:

Body resistance (r_b) decreased significantly from 334.5 to 270.9 s m⁻¹ as wind speed was increased from 0.25 to 2.0 m s⁻¹ (regression analysis, F_1,39=15.7, P=0.0003) (Fig. 4). A greater relative decrease was observed in the thermal resistance of the aerodynamic boundary layer (r_e), which declined by 64 % from 70 to 25 s m⁻¹ over the same range of wind speeds. Addition of body and boundary layer resistances (r_b+r_e), yielded total resistance (r_T) to heat flow between the animal and environment. r_T showed an overall decline of 27 % from 404.5 to 295.8 s m⁻¹ as wind speed increased from 0.25 to 2.0 m s⁻¹.

Body temperatures of birds did not vary systematically with wind speed, but were significantly affected by irradiance level at all wind speeds except 1.0 m s⁻¹ (Fig. 1). Body temperatures were typically highest under the conditions of greatest cold stress (e.g. in the absence of solar radiation). We cannot explain why, in general, the physical mechanism causing the changes in thermal stress (i.e. wind speed versus irradiance level) had different affects on T_b. Webster and Weathers (1988) found no correlation below 30 °C between T_b of verdins and wind speed, although T_b did increase with increasing air temperature T_a. Similar results were reported by Goldstein (1983) for Gambel’s quail (Callipepla gambelii).

Paladino and King (1984) found that exercising white-crowned sparrows increased their body temperature as work load and rates of oxygen consumption increased at a given temperature. In the same birds at rest, body temperatures increased erratically with declining air temperature and increasing rates of oxygen consumption. The available evidence suggests that T_b is unaffected by changes in the convective environment, but our data suggest that specific mechanisms of heat transfer (i.e. convection versus radiation) have differing effects on T_b.

Our measurements of metabolic heat production for white-crowned sparrows in the absence of simulated solar radiation are 18–28 % higher that those reported by Robinson et al. (1976) over a similar range of wind speeds and air temperatures. Differences in experimental conditions among these studies largely explain this variation. The measurements of Robinson et al. (1976) were made on birds during the rest phase of their daily cycle and would therefore be expected to be 20–25 % lower than those measured during the animal’s active phase (Aschoff and Pohl, 1970; Calder and King, 1974).

In addition, the birds of Robinson et al. (1976) were in a post-absorptive condition and held in a darkened wind tunnel for 3 h before any measurements were started. In our study, the birds were fed and held in the light for approximately 55 min before measurements. Furthermore, the birds in our study were blindfolded during trials so they perched quietly in the illuminated chamber. We found that normal room illumination (<3W m⁻²) in non-blindfolded birds resulted in an approximately 10 % increase in metabolic heat production above that in darkened conditions; all treatments were therefore carried out under some illumination (e.g. <3 or 936 W m⁻²). Published resting metabolic rates for white-crowned sparrows at 10 °C under free-convection conditions are variable, with DeJong (1976) reporting 106 W m⁻² for a post-absorptive bird under room lighting during the active phase of its daily cycle and King (1964) reporting 85 W m⁻² for post-absorptive birds held in the dark during their rest phase. Extrapolation of our data (Fig. 2; equation 7) to zero wind speed in the absence of short-wave radiation yields an estimated metabolic rate of 103.7 W m⁻².

As expected, thermal insulation decreased with increasing wind speed (Fig. 4); our values are lower than those reported for white-crowned sparrows by Robinson et al. (1976). Body resistances are calculated from rates of metabolic heat production and, as discussed above, rates of metabolic heat production reported in this study are 18–28 % higher than those reported by Robinson et al. (1976). Body thermal resistance values from the present study reflect these differences and are approximately 13 % lower than those reported by Robinson et al. (1976). However, sparrows from both studies showed very similar relative changes in r_b with increasing wind speed; as wind speed was increased from 0.5 to 2.0 m s⁻¹, the r_b of sparrows in the present study declined by approximately 19 %. Robinson et al. (1976) reported a 15 % decline in r_b under similar conditions. The total thermal resistance (r_T=r_b+r_E) of the animals in this study declined by 27 % from 404.5 to 295.8 s m⁻¹ as wind speed increased from 0.25 to 2.0 m s⁻¹.

The net solar heat gain (SHG) measured in white-crowned sparrows during this study declined by 56 % as wind speed was increased from 0.25 to 2.0 m s⁻¹ (Fig. 3). In a previous study (Wolf and Walsberg, 1996), net solar heat gain in verdins was found to be even more sensitive to changes in wind speed; SHG declined by more than 90 % as wind speed was increased from 0.4 to 3.0 m s⁻¹. Extrapolating from the white-crowned sparrow data to zero wind speed, solar heat gain is estimated to equal 32W m⁻² or 63 % of the sparrow’s basal metabolic rate (BMR=51 W m⁻²) reported by King (1964). This value is similar to the solar heat gain value reported by DeJong (1976) of 38.4 W m⁻² for white-crowned sparrows held at 10 °C and an irradiance of 1047 W m⁻² under free-convective conditions.

At 2.0 m s⁻¹, solar heat gain declined substantially and equaled 24 % of BMR.

Previously, we found that verdins experience large heat loads from solar radiation and that a very large proportion of the intercepted radiation is transferred to the skin and is seen as a reduction in metabolic rate (Wolf and Walsberg, 1996). If we assume the same shape factor used for the present study (0.28) (previously, and in error, we used a shape factor of 0.44) and apply the above equation to the verdin data for a wind speed of 0.4 m s⁻¹ and an irradiance of 500 W m⁻², we find that approximately 44 % of the intercepted radiation is used physiologically and is seen as a heat load to the skin. DeJong’s (1976) data suggest that the large disparity in the percentage of intercepted radiation transferred to the skin in verdins versus white-crowned sparrows and expressed as SHG may largely be due to the different irradiance levels used in the two studies. In the present study, sparrows were exposed to an irradiance of 936 W m⁻²; previously, we exposed verdins to an irradiance of 500 W m⁻² (Wolf and Walsberg, 1996). A reappraisal of DeJong’s (1976) data indicate that solar heat gain does not scale directly with irradiance level (Fig. 5), but is a negative power function of irradiance level. DeJong’s (1976) data for white-crowned sparrows suggest that doubling irradiance from 500 to 1000 W m⁻² in the absence of wind increases solar heat gain by only approximately 50 %. Walsberg et al. (1997) reported very similar results for a small mammal in a direct test of the effect of irradiance level on solar heat gain. As suggested by Walsberg et al. (1997), increased irradiance levels are likely to alter the heat balance of the system. Because heat loss through an insulating coat or plumage occurs by the coupled processes of long-wave radiation and free and forced convection, changes in irradiance level may alter the balance between these processes. Increased irradiance may result in increased penetration of radiation into the plumage and may therefore increase the thermal gradients within the plumage that drive heat flow. This could result in a decrease in plumage insulation and a concurrent reduction in solar heat gain.

In this study, we have examined how complex radiative and convective environments affect thermoregulatory costs and heat balance in a small bird. Body temperature is affected by irradiance level, but not by wind speed. Thus, different types of thermal loads have different effects on the animal’s body temperature. Wind speed had marked effects on solar heat gain. Increases in wind speed from 0.25 to 2.0 m s⁻¹ reduced SHG by approximately 56 %. This sensitivity of radiant heat loads to changes in the convective environment appears quite variable interspecifically, as shown by our previous study (Wolf and Walsberg, 1996). We also found significant variability in the proportion of intercepted radiation that affected the animal’s thermal balance. This varies both with wind speed and with the intensity of irradiance. For example, a fourfold increase in irradiance results in only an approximately 50 % increase in metabolic savings to the animal (calculated from DeJong, 1976). That is, low-intensity irradiance has much greater proportional effects on the animal’s heat balance than do high levels of irradiance. Such phenomena add a new layer of complexity to efforts to predict the effects of complex radiative and convective environments on the heat balance and energetics of animals in nature.

We thank Ty C. M. Hoffman for his untiring efforts to facilitate our research goals in the dark face of adversity. We also thank two anonymous reviewers, who provided very thorough reviews and thoughtful comments that markedly improved this paper, and Todd McWhorter for his comments on an earlier draft of this manuscript. This research was supported by NSF grants BSR 91-07470, IBN 92-24130 and IBN 96-27541 and The University of Arizona.